cjwatsonLaney: Did you have plans for dealing with haskell-platform?  7.4 -> 7.6 is a fairly big change, but it looks like the next upstream platform release is May, so we can't really wait00:21
cjwatsonLaney: Maybe it's best to just not claim to provide the platform when we can't ...00:25
ScottKinfinity or slangasek: If you're around ^^^ is a straightforward backport from upstream that it'd be nice to get in soonish.03:19
infinityScottK: Will look when launchpad gets diffy with it.03:20
ScottKinfinity: Thanks.  Diff can also be found at http://bazaar.launchpad.net/~kubuntu-packagers/kubuntu-packaging/kdepim/diff/260 if you'd rather not wait.03:21
infinityScottK: I don't trust reviewed branches to match packages. :P03:22
infinity(If I'm impatient, I could download and diff, but I'm in no rush)03:22
seb128hey release friends07:31
seb128could anyone review the indicator stack in the queue?07:31
seb128(I hoped it would have happened over night and be on today's iso but they are still sitting in there :-()07:31
Laneycjwatson: I think that could go away and be backported when it appears if we don't see release candidates from upstream08:07
Laneyimpressive effort debugging btw08:08
seb128Laney, \o/08:12
cjwatsonhmph, not good enough yet to get haskell-conduit/armhf building :-/08:14
cjwatson### Error in Data/Conduit.hs:8: expression `runResourceT $ sourceFile "input.txt" $$ sinkFile "output.txt"'08:14
cjwatsonfd:11: hGetLine: end of file08:14
cjwatsondoctests: fd:10: hClose: resource vanished (Broken pipe)08:14
cjwatsonTest suite doctests: FAIL08:14
cjwatsonwhich at least is different ...08:14
seb128whoever is reviewing indicators, thanks ;-)08:18
slangasekseb128: do you know what the story is with 'friends'?  It's an autolanded package with no bugs mentioned at all in the changelog; I'm not confident this is really meant to be going in post-freeze08:25
seb128slangasek, look at https://code.launchpad.net/~super-friends/friends/trunk08:26
seb128slangasek, you have the commits between the bot tags08:26
seb128slangasek, I assume that all the stuff autolanding are landing from the right (stable) vcs08:27
slangasekseb128: so why did we not get useful changelog entries?08:31
slangasekthese autoland branches are already a pain wrt queue reviews, without me having to hunt down branches besides08:33
slangasekand really, the commit messages there aren't any more enlightening - I don't see why these are freeze-appropriate changes, it looks like a refactor with no justification given08:33
slangasekrobru: ^^ these seem to be your changes, perhaps you could comment why this should go into raring this late in the cycle?08:34
LaneyYou get a changelog entry if the branch is linked to a bug, AIUI. I've previously argued that it should fall back the commit message rather than insert no message, but I didn't win that one.08:34
slangasekseb128: anyway, all done except for friends08:35
seb128slangasek, you have 2 options to get a changelog entry08:35
seb128- edit the changelog with your commit08:35
seb128- link to a bug08:35
seb128(either with commit --fixes or by adding a bug reference like (lp: #...)08:35
Davieyit does seem crackful to maintain a changelog file, for something which is stored in VCS and built automatically.. but what do i know :)08:36
seb128slangasek, but I agree, that commit/update seems like weakly justified08:36
seb128it's more a problem with the person who did the commit that with the system though08:36
seb128slangasek, thanks for reviewing the other ones!08:36
seb128Laney, issue with "fall back to commit message" is that if your project has translation commits you could end up with 15 "updated translations" entries in the changelog08:37
seb128and nothing else08:37
slangasekDaviey: I don't care how the information in the changelog is generated, I just insist that there should be some - in this case it seems the missing changelog is caused by them not being bugfixes08:37
seb128which is not the end of the world, but suboptimal as well08:37
Davieyseb128: I thought the recommended way of handling translations was having LP commit to a separate branch, and a single merge commit - when it is suitable for the maintainer of trunk?08:38
didrocksthanks slangasek for looking at them ;)08:39
didrocksLaney: see what seb128's told. If you have any commit message, I think this is making a lot of noise for nothing08:39
xnoxseb128: I remember writting hooks that strip "translation commits" but leave the rest in.08:39
didrocksLaney: but I'm happy to revisit this decision if needed at the sprint/UDS08:39
seb128Daviey, I'm not sure what's the "recommended" way, but see e.g http://bazaar.launchpad.net/~lightdm-team/lightdm/trunk/changes/1572?start_revid=163208:40
slangasekdidrocks: hey, just trying to clear the deck for my compiz fix to land ;)08:40
seb128Daviey, 1/3 of that page is "Launchpad automatic translations update."08:40
didrocksslangasek: ahah, it was interested ;)08:40
infinitydidrocks: Some commits may be "noise" (like "fixed typo from previous commit" or whatever), but there must be a clever way to tag "I want this commit message in the changelog" that isn't always "it must be a bug reference".08:40
didrocksslangasek: however, the integration tests for the unity stack failed today. I think mterry will have a look once around08:40
seb128infinity, you can manually add it to the changelog08:40
infinitydidrocks: Cause "I fixed this nasty bug or addeed this cool new feature before anyone noticed or asked for it" is pretty valid for a changelog. :P08:41
infinityseb128: Okay, but I suspect that feels like duplicate effort, which is why it's not happening.08:41
didrocksinfinity: yeah, what seb128 told, that's why you can reference directly to the changelog08:41
seb128infinity, suggesting of smarter way to flag commits as "should show in the changelog" are welcome I guess ;-)08:42
didrocksinfinity: I doubt that they will use a tag if they don't care about linking to a bug or adding a changelog entry TBH08:42
seb128we could add tags in the commit message08:42
seb128but that would be weird08:42
slangasekdidrocks: bah, thwarted at every turn08:42
seb128that would add "noise" in the history log08:42
infinityWorks for the kernel team.08:42
Laneythat's what git-dch has you do08:42
slangasekdidrocks: oh that reminds me, why is compiz's test suite not run at package build time?08:42
didrocksslangasek: probably a leftover of when it didn't run without X (they didn't separate that part). I'll just have a try in a pbuilder right now08:44
slangasekdidrocks: ah, ok.  I'm pretty sure it doesn't need X now, it was running fine for me here08:44
didrocksslangasek: great, let me have a try quickly ;)08:44
didrockswell quickly == as fast as compiz will build ;)08:45
Davieyseb128: Just tried to find where it is documented, and the only references i can find from an official source is because of it's forceful overwriting - which seems weak.08:46
didrocksLaney: infinity: let's discuss that during the sprint, having upstream on board and knowing why they don't do dch -i or link to a bug report will be useful08:47
didrocksLaney: infinity: and seeing if another workflow will be easer for them to respect for important changes08:47
didrocksLaney: infinity: btw, I wrote that and linked them to https://wiki.ubuntu.com/DailyRelease/FAQ#My_name_deserves_to_be_in_the_changelog.21 a while ago08:47
LaneySure. I just see empty changelogs fairly often and it's something that I'd like to work to avoid. I know that the process is supposed to block if this happens, but that doesn't always seem to work for whatever reason. :)08:48
cjwatsonAnother suggestion: fall back to commit message, but have a way to tag commits as not to be included in the changelog (e.g. "[SILENT]" in the commit message or something)08:48
didrockscjwatson: ah, so on by default and off on demand?08:49
didrockscjwatson: TBH, we tried tag to UNBLOCK merge when being in a freeze a year ago08:49
LaneyRight, with git-dch you put Git-Dch: Ignore in your changelog and it's skipped08:49
didrocksand most of the time, they forgot to add it08:49
cjwatsonSeems like that would address the "too much noise" problem in the case where it actually bothers people08:49
didrocksI'm afraid the same thing will happen in a commit08:49
cjwatsonYou could for example turn it off centrally for translation processing08:50
infinityIf a trivial commit slips into the changelog, that's not world-ending.08:50
infinityThe inverse is annoying.08:50
Laneynot that it's the end of the world if a few extra entries are added08:50
didrockswe can maybe try that, I want first to talk with them about it08:50
didrocksand see what's the best way to get in08:51
LaneyThat or have the merger reject stuff if it's got no commit message08:51
didrocksLaney: it's already the case08:51
* infinity decides to find a bed.08:51
didrocksinfinity: have a good night :)08:51
Laneyah, this one was a direct commit to the trunk branch08:52
Laneyis that what shouldn't be done?08:52
infinityWell, I was going to watch some TV before bed, but pulseaudio now hates me.08:52
didrocksslangasek: how did you run the tests? I'm getting:08:52
didrocks/usr/bin/ctest --force-new-ctest-process -j108:52
didrocksTest project /tmp/buildd/compiz-0.9.9~daily13.02.28/obj-x86_64-linux-gnu08:52
didrocksNo tests were found!!!08:52
didrocksLaney: it should never be done, right08:52
didrocksrobru: kenvandine ^08:53
slangasekdidrocks: ah, did you add the missing build-dep on libxorg-gtest-dev?08:53
slangasekdidrocks: (and enable it in the configure arguments?)08:53
didrocksslangasek: yeah, it's installed08:53
didrocksah, there is a enabling08:53
didrocksok, found it08:53
didrocksslangasek: hum, tests enabled, but nothing. I need to look closer when I've time for it09:04
=== doko_ is now known as doko
=== greyback is now known as greyback|lunch
=== greyback|lunch is now known as greyback
apwinfinity, ^^ is the ftbfs fix for the config issue you spotted, process changed to cover for next time13:01
kenvandinedidrocks, slangasek: i'll make sure merges get rejected without changelog entries or bug references, sorry about that14:08
kenvandineslangasek, those fixes are pretty important14:08
xnoxkenvandine: didrocks: maybe jenkins should reject branches without bug/ref and/or comments in debian/changelog for "stable" branches (post FFe & for SRU)14:28
kenvandinexnox, didrocks i'd be fine with that14:42
Laneydidrocks said earlier that it does that14:43
Laneybut this was a direct commit to trunk so no merger was involved14:43
kenvandinehumm.. the fixes i cared about weren't directly committed, but maybe there was some fix direct to trunk14:49
kenvandineoh, last night14:49
kenvandinerobru pushed the fix for python compatibility for quantal directly to trunk..14:50
kenvandineohhhhh.... he's been merging himself after the branches get reviewed14:51
kenvandineinstead of waiting for the merger...14:51
seb128Laney, no, the merge reject merge requests without a commit message14:51
kenvandinerobru, lets not do that :)14:51
seb128Laney, there is no requirement for bug reference or debian/changelog entry14:51
seb128oh maybe I'm wrong14:51
kenvandineindeed, so that wouldn't help here14:51
kenvandinei don't think there is anything that rejects without the changelog entry14:52
seb128right, what I though14:52
Laneyah ok, well that would be useful then14:52
kenvandinei've seen plenty of those uploaded with just a snapshot entry14:52
didrockskenvandine: it's because upstream is not respecting the deal: for things important, it's either a bug linked or a direct changelog update14:52
Laneysome other merger we have definitely does do that14:52
didrockskenvandine: I don't think you want to have every commits listed14:53
LaneyI've had to go back and set commit message before before a merge would be done by a bot14:53
didrocksand I'm not talking about a feature being merged with hundreds of commits :)14:53
kenvandinedidrocks, then it's hard to decide when there should be something listed14:53
didrocksLaney: yeah, this is the commit message for bzr14:53
LaneyI mean one on the MP14:53
Laneywhich is probably the right level to do it at?14:53
kenvandinethese were all small bug fixes14:53
didrockskenvandine: you would prefer to have everything listed?14:53
kenvandinebut not LP bugs for them14:53
Laneyanyway, back to what we said would be a topic for the sprint :P14:54
kenvandineso the issue is when it is optional,  it'll often get left out14:54
kenvandinei know our upstreams :)14:54
kenvandinei do really want the friends update that is in the queue to get through, fixes the dee cache :)14:55
kenvandinedidrocks, i just don't like uploads to the archive with a one line changelog entry  " * Automatic snapshot from revision 179"14:56
Laneyis there a convenient way to fetch the source of copies in the queue? queue can't do it.14:58
cjwatsonMight be worth just teaching queue how14:59
* Laney nods14:59
didrockskenvandine: well, that's because of upstream basically, it's like "I don't like upstream not following our guidelines" :)14:59
didrockskenvandine: you should lead by example and not have that on friends for instance :p14:59
kenvandineyeah... i know :)14:59
kenvandinebut you know... none of those were my changes :)14:59
kenvandinebut i did review them... i should have rejected :)15:00
didrocksright, we have 2 people to ack the change :)15:02
didrocksI can add the commit message when no bugs is linked15:02
didrocksbut I want upstream to agree first15:02
didrockslet's see during the sprint15:03
=== didrocks1 is now known as didrocks
=== jbicha_ is now known as jbicha
dokofinally, cython didn't fail for random reasons on armhf ...15:40
ScottKLaney: dgetlp mostly works if you have the path to othe .dsc.16:23
LaneyScottK: Yep, but that's the other side of 'convenient' to me.16:23
LaneyApparently package_upload objects which are copies don't have a reference to the originating archive16:24
ScottKApparently adding branding is a bug fix (ngnix)16:26
ScottKerr nginx16:26
LaneyBug: there is no branding16:28
LaneyFix: add branding16:28
ScottKI don't think adding (Ubuntu) to the version string for a web server is a great idea.  I'm going to reject it.16:28
jbichaScottK: apache on Ubuntu has done that for quite a while, see http://cdimage.ubuntu.com/404 for instance16:35
ScottKRight, but there are also some significant differences in Debian/Ubuntu on where data files are located and such for Apache, so it's actually relevant to know for troubleshooting.16:36
ScottKI don't think that's true for nginx, but whatever.16:36
ScottKIf Daviey wants it, he can accept it.16:36
rtg_infinity, cjwatson: please accept Raring linux-meta with Highbank Q->R upgrade fixes. Vanhoof has the wherewithal to test.16:42
antarusall hail cjwatson !17:14
dobeycan i bug anyone to look at a couple of UIFe requests please?17:44
infinityScottK: Differences between Debian and Ubuntu apache packaging?  Back when I maintained it, the only difference was the branding. :P17:45
ScottKinfinity: Sorry.  I meant Debian/Ubuntu different from the rest of the world, not from each other.17:46
infinity(And the point of the branding is entirely to hint people who like to do silly "which OS is more popular" statistics gathering).17:46
slangasekdobey: bugs numbers?17:46
ScottKinfinity: I won't object if someone wants to accept it from rejected.17:46
dobeyslangasek: bug #1151621 and bug #116635617:46
ubot2Launchpad bug 1151621 in Ubuntu Software Center stable-5-6 "[UIFe] TypeError when opening edit menu" [Undecided,Triaged] https://launchpad.net/bugs/115162117:46
ubot2Launchpad bug 1166356 in rhythmbox-ubuntuone (Ubuntu Quantal) "[UIFe] Old music store interface going away on server" [Undecided,New] https://launchpad.net/bugs/116635617:46
hallynis now a bad time to push a small lxc fix for bug 1166870?18:13
ubot2Launchpad bug 1166870 in lxc (Ubuntu) "lxc-clone fails silently" [High,In progress] https://launchpad.net/bugs/116687018:13
hallyn(FinalBetaFreeze link on https://wiki.ubuntu.com/RaringRingtail/ReleaseSchedule leads to nonexistant wiki page018:17
hallynoh, should final beta freeze be taken out of topic, as final beta release should have happened?18:18
=== slangasek changed the topic of #ubuntu-release to: Ubuntu 12.10 and 12.04.2 released | Archive: Frozen for release | Raring Ringtail Release Coordination. Please don't upload things during freezes where you shouldn't, or be prepared to apologise to the release team | we accept payment in cash, check or beer | melior malum quod cognoscis
slangasekhallyn: done18:19
hallynslangasek: so i guess that's a no about uploading lxc? :)18:19
slangasekhallyn: "frozen" in the sense of "gated", not "immobile"18:20
slangasekbugfix uploads are still accepted18:20
hallynok thx18:20
slangasekdobey: I don't understand from bug #1151621 what you're requesting a UIFe on18:27
ubot2Launchpad bug 1151621 in Ubuntu Software Center stable-5-6 "[UIFe] TypeError when opening edit menu" [Undecided,Triaged] https://launchpad.net/bugs/115162118:27
slangasekdobey: UI Freeze governs changes to text strings and window layout that would invalidate documentation, translations, screenshots - fixing a UI bug doesn't need a UIFe18:28
infinityWell, introducing a UI bug to work around a crash.  But yeah.18:29
infinityIf there are screenshots of people pulling down the Edit menu in software-center, I'd be surprised.18:29
dobeyslangasek: i don't know if it breaks any docs or not. hence the request for a UIFe for people who own docs/translations to say that it doesn't.18:30
dobeyslangasek: my understanding was that all UI changes needed to request freeze exceptions at this point18:31
slangasekdobey: and bug #1166356 seems to be about an SRU rather than an upload to raring that needs a UIFe?  You've marked the raring task 'fix released'18:31
ubot2Launchpad bug 1166356 in rhythmbox-ubuntuone (Ubuntu Quantal) "[UIFe] Old music store interface going away on server" [Undecided,New] https://launchpad.net/bugs/116635618:31
dobeyslangasek: yes, the new version is raring, i want to SRU the new version to quantal/precise18:31
slangasekdobey: well, you haven't given any information in that bug to /tell/ the docs people what's changing that might impact them18:31
slangasekI don't see how a crash fix has any impact on docs18:32
dobeyslangasek: because generally docs tend to document the UI, the edit menu of which is one aspect of said UI18:32
infinitydobey: I appreciate that you're being careful and cautious here.  In this case, probably a bit too much so, that's all. :)18:33
infinitydobey: If we have application specific docs that tell users all the ways to copy/paste, those docs are probably wrong. :P18:33
slangasekdobey: ok, on reread I see that the question is about removing one of the menu entries - right, that's reasonable18:33
slangasekdobey: UIFe approved18:33
dobeyand the rhythmbox-ubuntuone one is about removing the UI on quantal/precise as well (it's already gone in raring)18:34
slangasekdobey: yeah, I think that's fine.. invalidating any screenshots of a UI that needs to go away is the lesser evil18:34
dobeyyeah, i suppose we'll need to drop the screenshot from the 12.04 installer for the next point release18:35
dobeyshould i add ubiquity to that bug report?18:35
slangasekdobey: if there's a screenshot of this in the installer, it'd be in ubiquity-slideshow-ubuntu18:38
slangasekballoons, tjaalton; tjaalton, balloons21:01
tjaaltonballoons: yo21:02
tjaaltonballoons: so, I'd need more people to test a new mesa release21:02
tjaaltonwhich has a FFE bug open21:02
tjaaltonballoons: I sent a CFT to ubuntu-devel@ last friday, but it didn't attract that many people, or they are shy to report success/failure to LP21:05
balloonstjaalton, hello :-)21:08
balloonswe can certainly organize something for you21:08
tjaaltonballoons: so it's for raring, and available on ppa:ubuntu-x-swat/x-staging21:09
slangasekballoons: basically, I've told tjaalton just now on #ubuntu-devel that I'm not comfortable accepting mesa in as a freeze exception without some wider testing; do you think it's realistic to gather info on it by next Tue at the latest?21:11
slangasekwe would need assurance that a wide assortment of GPUs still work without regression21:11
slangasekI would say we'd want evidence from at least 4 different GPU flavors in each of the three families21:12
tjaaltonsomething like that21:12
balloonsslangasek, we can try -- now is a quite busy time of course, but a week should be enough time to get people testing. 4 different gpu flavors from each of intel, amd, and nvidia yes?21:12
tjaaltonwe already have piglit testing done21:12
slangasekballoons: yse21:12
balloonshas this gone into the lab?21:12
balloonsthat's your best bet at the moment to get specific testing on different gpus21:13
slangasekplus some specific spot-checking for regressions on ARM (LLVM?), and spot-checking that it doesn't regress anything in combination with the binary drivers21:13
tjaaltonllvm is still a no-go21:13
tjaaltonon arm21:13
balloonsarm is broken..21:13
ogra_arm image should be fine again and you can at least test swrast21:14
balloonsogra_, did you fix it?21:14
balloonsmaybe I wasn't subbed on the bug, i didn't see anything about a fix21:14
ogra_nope, mlankhorst did21:15
ogra_you mean the "black screen after install" one, right ?21:15
ogra_yup, a fix went in for that one21:15
slangasektjaalton: well, llvm is "too slow to be worth anything for unity" on ARM because it's not ported to NEON... but someone should check in some capacity that LLVM isn't broken21:16
ogra_bug 116198121:16
ubot2Launchpad bug 1161981 in xorg-server (Ubuntu) "Boot stalls after Ubuntu Raring desktop ARM (Panda board) install" [High,Fix released] https://launchpad.net/bugs/116198121:16
slangasekmaybe just an x86 VM test?21:16
ogra_balloons, ^^^21:16
ogra_went in yesterday ... should be fine on todays images21:16
balloonsogra_, ty.. must not have subbed :-)21:16
tjaaltonslangasek: sure thing21:16
tjaaltonI have a speedy panda for it..21:17
ogra_speedy ...21:17
* ogra_ pets his chromebook21:17
tjaaltonand nexus7 of course21:17
ogra_yeah, thats for the binary drivers :)21:17
slangasektjaalton: that sound like enough testing to you? :)21:18
tjaaltonslangasek: yeah21:18
balloonsok, so I missed the response.. Has this been run in the lab tjaalton ?21:19
tjaaltonballoons: nope21:19
balloonstjaalton, ok, so let's make that happen.. that will be the best thing to do to make sure it passes Steve's feel good test.. and after all you need him to feel good ;-)21:22
balloonsalongside we'll pull community results to21:22
tjaaltonsounds like a plan :)21:23
slangasekis the lab testing something we should be doing more systematically wrt mesa uploads?21:23
slangasekseems like a good candidate package for "critical ppa" pre-acceptance testing21:23
slangasekto at least give you X guys solid data about any possible regressions21:24
balloonsslangasek, I think that's something worth chatting with the qa team about.21:25
balloonshave them track the xorg-edgers crack ppa or something :-)21:26
slangasekthis is indeed something that should happen21:26
balloonswhere's the ffe link btw?21:26
slangasekI'm surprised that's not already a priority for the lab, but I guess unity regression-testing is more critical21:26
tjaaltonhmm actually there's a graphics test suite of some sorts they should be running when certifying new platforms21:26
tjaaltonor is this the same lab we're talking about?21:26
tjaaltonballoons: https://launchpad.net/bugs/116409321:27
ubot2Launchpad bug 1164093 in mesa (Ubuntu) "FFe: new upstream release 9.1.1" [Wishlist,Triaged]21:27
slangasektjaalton: different lab21:27
tjaaltonone thing I know the new version fixes is user-switching no longer hangs compiz on intel, like it sometimes does on mesa 9.0.x :)21:44
=== Ursinha-afk is now known as Ursinha
=== elmo__ is now known as elmo
slangasekcyphermox_: huh, so this nm upload only changes the test suite?22:24
slangasekI guess that's a pretty safe change22:24
cyphermox_slangasek: yeah, only adds tests23:35

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!